Method, device, equipment and medium for predicting radio wave propagation loss
Through the preset loss prediction model, combined with urban images and radio wave transmitter parameters, the electromagnetic situation chart is predicted, which solves the prediction problem of radio wave propagation loss in complex urban environments, and provides decision-making reference for parameter selection of vehicle network communication systems.
Patent Information
- Application Number
- CN202411001930.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-25
- Publication Date
- 2024-11-26
AI Technical Summary
The prior art lacks effective solutions to predict radio wave propagation losses in complex urban environments, affecting the parameter selection of Internet of Vehicles communication systems.
By obtaining target data, including urban images and radio wave transmitter parameters, input it to the preset loss prediction model, output the predicted electromagnetic situation chart, and predicting radio wave propagation loss using the preset loss prediction model based on multiple sets of training data.
Effective prediction of radio wave propagation loss is achieved, the gap in the existing technology is filled, and the ability to reasonably predict radio wave propagation loss is achieved.
Smart Images

Figure CN119030644A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of radio wave propagation technology, and in particular to a method, device, electronic device, non-transitory computer-readable storage medium and computer program product for predicting radio wave propagation loss. Background Art
[0002] In recent years, with the evolution of new generation information technologies such as artificial intelligence, mobile communication systems and Internet of Vehicles communication technologies have developed rapidly. Among them, millimeter wave communication technology is the key and core to support the Internet of Vehicles system (vehicle positioning and location perception). Therefore, it is urgent to study the prediction of communication scenario status in complex urban environments based on the current application scenarios of the Internet of Vehicles, so as to provide decision-making references for the parameter selection of the communication system. Among them, the prediction of communication scenario status involves the prediction of radio wave propagation loss.
[0003] However, there is currently no mature solution for predicting radio wave propagation losses. Summary of the invention
[0004] The present invention provides a method, an apparatus, an electronic device, a non-transitory computer-readable storage medium and a computer program product for predicting radio wave propagation loss, so as to solve the defects in the prior art.
[0005] The present invention provides a method for predicting radio wave propagation loss, comprising: Acquire target data to be processed; wherein the target data includes an urban area image associated with the target urban area and radio wave transmitter parameters, and the urban area image is a cross-sectional image of the three-dimensional model of the target urban area at a predetermined height; Inputting the target data into a preset loss prediction model, and outputting a predicted electromagnetic situation map via the preset loss prediction model as a prediction result of the radio wave propagation loss; Among them, the preset loss prediction model is trained based on at least one set of training data pairs, each set of training data pairs in the at least one set of training data pairs includes a sample input and a sample true value, the sample input includes a sample urban area image and a sample radio wave transmitter parameter, and the sample true value includes an electromagnetic situation map corresponding to the sample input.
[0006] The present invention also provides a device for predicting radio wave propagation loss, comprising: The data acquisition module is configured to: acquire target data to be processed; wherein the target data includes an urban area image associated with the target urban area and radio wave transmitter parameters, and the urban area image is a cross-sectional image of the three-dimensional model of the target urban area at a predetermined height; The loss prediction module is configured to: input the target data into a preset loss prediction model, and output a predicted electromagnetic situation map via the preset loss prediction model as a prediction result of the radio wave propagation loss; Among them, the preset loss prediction model is trained based on at least one set of training data pairs, each set of training data pairs in the at least one set of training data pairs includes a sample input and a sample true value, the sample input includes a sample urban area image and a sample radio wave transmitter parameter, and the sample true value includes an electromagnetic situation map corresponding to the sample input.
[0007] The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the method for predicting radio wave propagation loss as described in any one of the above is implemented.
[0008] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements any of the above-described methods for predicting radio wave propagation loss.
[0009] The present invention also provides a computer program product, comprising a computer program, wherein when the computer program is executed by a processor, the computer program implements any of the above methods for predicting radio wave propagation loss.
[0010] As described above, the method for predicting radio wave propagation loss provided by the embodiment of the present invention can effectively predict radio wave propagation loss by utilizing the information integration capability of the preset loss prediction model, thereby filling the gap in the prior art. The preset loss prediction model is obtained by performing multi-dimensional training based on multiple sets of sample data under different experimental conditions, so it has the ability to reasonably predict radio wave propagation loss based on the input target data to be processed. BRIEF DESCRIPTION OF THE DRAWINGS
[0011] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.
[0012] Figure 1 This is one of the flow charts of the method for predicting radio wave propagation loss provided by an embodiment of the present invention; Figure 2 This is a second flow chart of a method for predicting radio wave propagation loss provided by an embodiment of the present invention; Figure 3 is a schematic diagram of the sample urban area image provided by an embodiment of the present invention; Figure 4 is a schematic diagram of the predicted electromagnetic situation map provided by an embodiment of the present invention; Figure 5 is a schematic diagram of the structure of the ResUNet-based neural network provided in an embodiment of the present invention; Figure 6 is a schematic diagram of the structure of a device for predicting radio wave propagation loss provided by an embodiment of the present invention; Figure 7 It is a structural schematic diagram of the electronic device provided by the present invention. DETAILED DESCRIPTION
[0013] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0014] In order to solve the above problems of the prior art, the inventors of the present invention have proposed a solution for predicting radio wave propagation loss as described in the following embodiment after research.
[0015] Example: The following describes the implementation details of the scheme for predicting radio wave propagation loss of the present invention in conjunction with the accompanying drawings.
[0016] Figure 1 FIG. 1 is a flow chart of a method for predicting radio wave propagation loss provided by an exemplary embodiment of the present invention. This embodiment can be applied to electronic devices (eg, servers or cloud computing platforms), such as Figure 1 As shown, the method for predicting radio wave propagation loss includes the following steps: S110: Obtain target data to be processed.
[0017] The target data includes an urban area image associated with the target urban area and radio wave transmitter parameters, and the urban area image is a cross-sectional image of the three-dimensional model of the target urban area at a predetermined height.
[0018] The present invention does not limit the "predetermined height", which can be determined according to actual needs or circumstances, for example, 50 meters or 100 meters from the ground from a bird's eye view.
[0019] The present invention does not limit the acquisition method. For example, an electronic device (eg, a server or a cloud computing platform) as an execution subject may communicate with a server storing the “target data to be processed” to acquire the “target data to be processed”.
[0020] S120, inputting the target data into a preset loss prediction model, and outputting a predicted electromagnetic situation map via the preset loss prediction model as a prediction result of the radio wave propagation loss.
[0021] Among them, the preset loss prediction model is trained based on at least one set of training data pairs, each set of training data pairs in the at least one set of training data pairs includes a sample input and a sample true value, the sample input includes a sample urban area image and a sample radio wave transmitter parameter, and the sample true value includes an electromagnetic situation map corresponding to the sample input.
[0022] Optionally, the radio wave transmitter parameters and sample radio wave transmitter parameters include position parameters, frequency parameters and power parameters. The physical property information of each pixel of the urban area image and the sample urban area image includes electromagnetic characteristic parameters of buildings, vegetation and terrain. The electromagnetic situation map and the predicted electromagnetic situation map contain multiple information, including received power, path loss, arrival angle, arrival time, delay spread and multipath number.
[0023] The structure and training steps of the preset loss prediction model will be described below and will not be elaborated here.
[0024] As described above, the method for predicting radio wave propagation loss provided by the embodiment of the present invention can effectively predict radio wave propagation loss by utilizing the information integration capability of the preset loss prediction model, thereby filling the gap in the prior art. The preset loss prediction model is obtained by performing multi-dimensional training based on multiple sets of sample data under different experimental conditions, so it has the ability to reasonably predict radio wave propagation loss based on the input target data to be processed.
[0025] exist Figure 1 Based on the embodiment, as an optional implementation method, refer to Figure 2 , the training steps of the preset loss prediction model include: S210: Obtain the at least one set of training data pairs.
[0026] As an optional example, S210 may be implemented in the following manner: 1) Based on the selected sample urban area, sample 3D model data (i.e., urban environment physical model data) is obtained.
[0027] The sample three-dimensional model data can be obtained in a variety of ways; for example, it can be collected from existing GIS systems, urban planning departments, public databases, or through LiDAR, drones and other equipment.
[0028] The sample dimensional model data may include building, vegetation and terrain information. Specifically, building information may include height, location, shape and material characteristics; vegetation information may include category and height; terrain information may include height and ground material characteristics.
[0029] 2) Preprocessing the sample three-dimensional model data obtained in 1). The preprocessing may include preprocessing the sample input and preprocessing the sample true value.
[0030] The preprocessing of sample input includes sample three-dimensional model conversion and sample radio wave transmitter parameter standardization. Specifically, sample three-dimensional model data conversion: the cross section of the sample three-dimensional model data at a specified height can be extracted as a two-dimensional image, while retaining the physical properties of each pixel (for example, electromagnetic properties of buildings, vegetation, and terrain); in addition, the sample three-dimensional model data needs to be cleaned and simplified as necessary to ensure data consistency while reducing computational complexity. Sample radio wave transmitter parameter standardization: the coordinates and related parameters of the sample radio wave transmitter can be standardized, the position coordinates can be normalized to the range of 0-1, and the transmission frequency and power can be normalized. Among them, after the above processing, the sample urban area image and sample radio wave transmitter parameters can be obtained.
[0031] The preprocessing of the sample truth value is the calculation of the electromagnetic situation map. Specifically, the ray tracing algorithm is used to simulate the electromagnetic situation under different environments and transmitter conditions to generate label data. Since a high-frequency ray tracing algorithm is used, the multipath information of each receiving point can be obtained, and all data can be obtained through processing, including receiving power, path loss, arrival angle, arrival time, delay spread, number of multipaths, etc. Among them, the electromagnetic situation map (as the sample truth value) can be obtained after the above processing.
[0032] In summary, through the above steps 1) to 2), the at least one set of training data pairs can be obtained.
[0033] Optionally, due to the limitation of the number of devices and computing power, it is usually difficult for the training data pairs to reach the number (in units of 10,000) required for model training. To solve this problem, in the embodiment of the present invention, random rotation, scaling, translation and other operations can be used to process the sample urban area images in the training data pairs to increase data diversity, thereby improving the generalization ability of the model trained based on the enhanced data.
[0034] S220, respectively, the sample urban area images (for example, Figure 3 The BuildingMap shown in FIG. 1 and the sample radio wave transmitter parameters are input into the loss prediction model to be trained, and the loss prediction model to be trained outputs the corresponding predicted electromagnetic situation map (for example, Figure 4The path loss coverage map shown is Coverage map of PathLoss).
[0035] Reference Figure 3 The sample urban area image shown, where white indicates there are buildings and black indicates there are no buildings. Figure 4 The predicted electromagnetic situation map is shown in Figure 1, where different colors (such as Figure 4 The red and blue areas shown represent different degrees of loss values, that is, electromagnetic situations.
[0036] S230. Using a preset comprehensive loss function, based on the electromagnetic situation map in each set of training data pairs and the corresponding predicted electromagnetic situation map, obtain a function value of the comprehensive loss function.
[0037] As an optional example, S230 may be implemented in the following manner: I1) Based on the electromagnetic situation map in each group of training data pairs and the corresponding predicted electromagnetic situation map, the function value of the received power loss function, the function value of the path loss function, the function value of the arrival angle loss function, the function value of the arrival time loss function, the function value of the delay spread loss function and the function value of the multipath number loss function are calculated respectively.
[0038] Optionally, the received power in the electromagnetic situation map in each set of training data pairs and the received power in the corresponding predicted electromagnetic situation map can be used to calculate L2 The norm is used as the function value of the receiving power loss function. Specifically, the following calculation formula (1) can be used to calculate the function value of the receiving power loss function: : (1) in, N is the number of the at least one set of training data pairs; Indicates i The received power in the electromagnetic situation map corresponding to the set of training data; Indicates i The received power in the electromagnetic situation map corresponding to the set of training data.
[0039] Similarly, the function value of the above-mentioned receiving power loss function can be referred to as The calculation method is to calculate the function value of the path loss function respectively. , the function value of the arrival angle loss function , the function value of the arrival time loss function , the function value of the delay spread loss function And the function value of the multipath number loss function .
[0040] I2) Using preset weights, weighted fusion is performed on the function values of the received power loss function, the function value of the path loss function, the function value of the arrival angle loss function, the function value of the arrival time loss function, the function value of the delay spread loss function, and the function value of the multipath number loss function to obtain the function value of the comprehensive loss function.
[0041] Specifically, the function value of the comprehensive loss function can be calculated using the following calculation formula (2): .
[0042] (2) in, to The weights of the received power loss function, the path loss function, the arrival angle loss function, the arrival time loss function, the delay spread loss function and the multipath number loss function can be represented in sequence. The weights can be determined according to the requirements, and the present invention does not limit this.
[0043] S240. Based on the function value of the comprehensive loss function, the loss prediction model to be trained is trained until a preset training completion condition is met, and the loss prediction model is obtained from the loss prediction model to be trained.
[0044] As an optional example, S240 may be implemented in the following manner: Passing back the function value of the comprehensive loss function to the loss prediction model to be trained; The loss prediction model to be trained adjusts the network parameters of each network layer according to the function value of the comprehensive loss function; Iterate the steps of obtaining the function value of the comprehensive loss function to adjusting the network parameters of each network layer until a preset training completion condition is met.
[0045] Among them, the preset training completion conditions include the number of training iterations reaching a preset iteration threshold, or the function value of the comprehensive loss function converges.
[0046] Based on the above embodiment, as an optional implementation, the preset loss prediction model is a neural network based on ResUNet.
[0047] Based on the above embodiment, as an optional implementation mode, refer to Figure 5 The neural network of the ResUNet includes a cascaded encoder unit, a dilated spatial pyramid pooling unit ASPP and a decoder unit.
[0048] The encoder unit includes a cascaded first encoding convolution module, a second encoding convolution module, a third encoding convolution module, a fourth encoding convolution module and a fifth encoding convolution module. The step length of the convolution layer in the first encoding convolution module is 1, the step length of the convolution layer in the second encoding convolution module is 1 / 2, the step length of the convolution layer in the third encoding convolution module is 1 / 4, the step length of the convolution layer in the fourth encoding convolution module is 1 / 8, and the step length of the convolution layer in the second encoding convolution module is 1 / 16.
[0049] It should be noted that the first coding convolution module, the second coding convolution module, the third coding convolution module, the fourth coding convolution module and the fifth coding convolution module may include residual blocks to extract and compress input features. The residual blocks alleviate the gradient vanishing problem and accelerate convergence by introducing shortcut connections.
[0050] The atrous spatial pyramid pooling unit ASPP consists of a cascade of multiple atrous convolutional layers with different sampling rates, a global average pooling layer, and a fully connected layer.
[0051] The decoder unit includes a cascaded first decoding convolution module and a second decoding convolution module; wherein the step size of the convolution layer in the first decoding convolution module is 1 / 4, and the step size of the convolution layer in the second decoding convolution module is 1.
[0052] An output end of the third encoding convolution module is jump-connected to an output end of the first decoding convolution module.
[0053] It should be noted that the features extracted by the encoder unit are gradually restored through upsampling and skip connections, and an electromagnetic situation map with the same size as the input is generated. Among them, the skip connection helps to retain high-resolution features and improve the prediction accuracy of the model.
[0054] In summary, in order to improve the generalization ability of the model, the ResUNet model in this embodiment uses a skip connection structure to ensure that multi-scale information can be captured by the model. The model performs convolution downsampling on the input data, then deconvolution upsampling, and fusion of the low-level feature maps before cropping.
[0055] As described above, the method for predicting radio wave propagation loss provided by the embodiment of the present invention can effectively predict radio wave propagation loss by utilizing the information integration capability of the preset loss prediction model, thereby filling the gap in the prior art. The preset loss prediction model is obtained by performing multi-dimensional training based on multiple sets of sample data under different experimental conditions, so it has the ability to reasonably predict radio wave propagation loss based on the input target data to be processed.
[0056] The following is a description of the device for predicting radio wave propagation loss provided by the present invention. The device for predicting radio wave propagation loss described below and the method for predicting radio wave propagation loss described above can be referenced to each other.
[0057] Figure 6 FIG. 1 is a schematic diagram of a device for predicting radio wave propagation loss provided by an exemplary embodiment of the present invention. Figure 6 As shown, the device for predicting radio wave propagation loss comprises: The data acquisition module 610 is configured to: acquire target data to be processed; wherein the target data includes an urban area image and radio wave transmitter parameters associated with the target urban area, and the urban area image is a cross-sectional image of the three-dimensional model of the target urban area at a predetermined height; the loss prediction module 620 is configured to: input the target data into a preset loss prediction model, and output a predicted electromagnetic situation map via the preset loss prediction model as a prediction result of the radio wave propagation loss; wherein the preset loss prediction model is trained based on at least one set of training data pairs, and each set of training data pairs in the at least one set of training data pairs includes a sample input and a sample true value, the sample number input includes a sample urban area image and a sample radio wave transmitter parameter, and the sample true value includes the electromagnetic situation map corresponding to the sample input.
[0058] Figure 7 An example of a physical structure diagram of an electronic device is shown in FIG. Figure 7 As shown, the electronic device may include: a processor 710, a communication interface 720, a memory 730 and a communication bus 740, wherein the processor 710, the communication interface 720 and the memory 730 communicate with each other through the communication bus 740. The processor 710 may call the logic instructions in the memory 730 to execute a method for predicting radio wave propagation loss, the method comprising: obtaining target data to be processed; wherein the target data comprises an urban area image associated with a target urban area and radio wave transmitter parameters, the urban area image being a cross-sectional image of a three-dimensional model of the target urban area at a predetermined height; inputting the target data into a preset loss prediction model, and outputting a predicted electromagnetic situation map via the preset loss prediction model as a prediction result of the radio wave propagation loss; wherein the preset loss prediction model is obtained by training based on at least one set of training data pairs, each set of training data pairs in the at least one set of training data pairs comprises a sample input and a sample true value, the sample number input comprises a sample urban area image and a sample radio wave transmitter parameter, and the sample true value comprises an electromagnetic situation map corresponding to the sample input.
[0059] In addition, the logic instructions in the above-mentioned memory 730 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on this understanding, the technical solution of the present invention can be essentially or partly embodied in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium, including several instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc. Various media that can store program codes.
[0060] On the other hand, the present invention also provides a computer program product, which includes a computer program, which can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the method for predicting radio wave propagation loss provided by the above methods, the method including: obtaining target data to be processed; wherein the target data includes an urban area image and radio wave transmitter parameters associated with a target urban area, and the urban area image is a cross-sectional image of a three-dimensional model of the target urban area at a predetermined height; inputting the target data into a preset loss prediction model, and outputting a predicted electromagnetic situation map through the preset loss prediction model as a prediction result of the radio wave propagation loss; wherein the preset loss prediction model is trained based on at least one set of training data pairs, each set of training data pairs in the at least one set of training data pairs includes a sample input and a sample true value, the sample number input includes a sample urban area image and a sample radio wave transmitter parameter, and the sample true value includes an electromagnetic situation map corresponding to the sample input.
[0061] On the other hand, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to execute the method for predicting radio wave propagation loss provided by the above-mentioned methods, the method comprising: obtaining target data to be processed; wherein the target data comprises an urban area image and radio wave transmitter parameters associated with a target urban area, the urban area image being a cross-sectional image of a three-dimensional model of the target urban area at a predetermined height; inputting the target data into a preset loss prediction model, and outputting a predicted electromagnetic situation map via the preset loss prediction model as a prediction result of the radio wave propagation loss; wherein the preset loss prediction model is trained based on at least one set of training data pairs, each set of training data pairs in the at least one set of training data pairs comprises a sample input and a sample true value, the sample number input comprises a sample urban area image and a sample radio wave transmitter parameter, and the sample true value comprises an electromagnetic situation map corresponding to the sample input.
[0062] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.
[0063] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.
[0064] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.
Claims
1. A method for predicting radio wave propagation loss, characterized in that: include: Acquire target data to be processed; wherein the target data includes an urban area image associated with the target urban area and radio wave transmitter parameters, and the urban area image is a cross-sectional image of the three-dimensional model of the target urban area at a predetermined height; Inputting the target data into a preset loss prediction model, and outputting a predicted electromagnetic situation map via the preset loss prediction model as a prediction result of the radio wave propagation loss; Among them, the preset loss prediction model is trained based on at least one set of training data pairs, each set of training data pairs in the at least one set of training data pairs includes a sample input and a sample true value, the sample input includes a sample urban area image and a sample radio wave transmitter parameter, and the sample true value includes an electromagnetic situation map corresponding to the sample input.
2. The method according to claim 1, characterized in that The radio wave transmitter parameters and sample radio wave transmitter parameters include position parameters, frequency parameters and power parameters; The physical attribute information of each pixel of the urban area image and the sample urban area image includes electromagnetic characteristic parameters of buildings, vegetation and terrain; The electromagnetic situation map and the predicted electromagnetic situation map contain a variety of information, including received power, path loss, arrival angle, arrival time, delay spread and multipath quantity.
3. The method according to claim 2, characterized in that The training steps of the preset loss prediction model include: Acquire the at least one set of training data pairs; The sample urban area image and the sample radio wave transmitter parameters in each set of training data pairs are respectively input into the loss prediction model to be trained, and the loss prediction model to be trained outputs the corresponding predicted electromagnetic situation map; Using a preset comprehensive loss function, based on the electromagnetic situation map in each set of training data pairs and the corresponding predicted electromagnetic situation map, a function value of the comprehensive loss function is obtained; Based on the function value of the comprehensive loss function, the loss prediction model to be trained is trained until a preset training completion condition is met, and the loss prediction model is obtained from the loss prediction model to be trained.
4. The method according to claim 3, characterized in that Using a preset comprehensive loss function, based on the electromagnetic situation map in each set of training data pairs and the corresponding predicted electromagnetic situation map, obtaining a function value of the comprehensive loss function includes: Based on the electromagnetic situation map in each set of training data pairs and the corresponding predicted electromagnetic situation map, respectively calculate the function value of the received power loss function, the function value of the path loss function, the function value of the arrival angle loss function, the function value of the arrival time loss function, the function value of the delay spread loss function and the function value of the multipath number loss function; Using preset weights, the function values of the receiving power loss function, the path loss function, the arrival angle loss function, the arrival time loss function, the delay spread loss function and the multipath number loss function are weightedly fused to obtain the function value of the comprehensive loss function.
5. The method according to claim 3, characterized in that: Based on the function value of the comprehensive loss function, the loss prediction model to be trained is trained until a preset training completion condition is met, including: Passing back the function value of the comprehensive loss function to the loss prediction model to be trained; The loss prediction model to be trained adjusts the network parameters of each network layer according to the function value of the comprehensive loss function; Iterate the steps of obtaining the function value of the comprehensive loss function to adjusting the network parameters of each network layer until a preset training completion condition is met.
6. The method according to claim 1, characterized in that The preset loss prediction model is a neural network based on ResUNet.
7. The method according to claim 6, characterized in that The neural network of the ResUNet includes a cascaded encoder unit, a dilated spatial pyramid pooling unit, and a decoder unit; The encoder unit includes a cascaded first encoding convolution module, a second encoding convolution module, a third encoding convolution module, a fourth encoding convolution module and a fifth encoding convolution module; wherein the step length of the convolution layer in the first encoding convolution module is 1, the step length of the convolution layer in the second encoding convolution module is 1 / 2, the step length of the convolution layer in the third encoding convolution module is 1 / 4, the step length of the convolution layer in the fourth encoding convolution module is 1 / 8, and the step length of the convolution layer in the second encoding convolution module is 1 / 16; The atrous spatial pyramid pooling unit includes a cascade of multiple atrous convolutional layers with different sampling rates, a global average pooling layer, and a fully connected layer; The decoder unit comprises a cascaded first decoding convolution module and a second decoding convolution module; wherein the step size of the convolution layer in the first decoding convolution module is 1 / 4, and the step size of the convolution layer in the second decoding convolution module is 1; An output end of the third encoding convolution module is jump-connected to an output end of the first decoding convolution module.
8. A device for predicting radio wave propagation loss, characterized in that: include: The data acquisition module is configured to: acquire target data to be processed; wherein the target data includes an urban area image associated with the target urban area and radio wave transmitter parameters, and the urban area image is a cross-sectional image of the three-dimensional model of the target urban area at a predetermined height; The loss prediction module is configured to: input the target data into a preset loss prediction model, and output a predicted electromagnetic situation map via the preset loss prediction model as a prediction result of the radio wave propagation loss; Among them, the preset loss prediction model is trained based on at least one set of training data pairs, each set of training data pairs in the at least one set of training data pairs includes a sample input and a sample true value, the sample input includes a sample urban area image and a sample radio wave transmitter parameter, and the sample true value includes an electromagnetic situation map corresponding to the sample input.
9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the program, the method for predicting radio wave propagation loss according to any one of claims 1 to 7 is implemented.
10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method for predicting radio wave propagation loss according to any one of claims 1 to 7 is implemented.
